NFSv4.1: Don't trust attributes if a pNFS LAYOUTCOMMIT is outstanding

If a LAYOUTCOMMIT is outstanding, then chances are that the metadata server may still be returning incorrect values for the change attribute, ctime, mtime and/or size. Just ignore those attributes for now, and wait for the LAYOUTCOMMIT rpc call to finish. Reported-by: shaobingqing <shaobingqing@bwstor.com.cn> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Intel-BMC · Jan 13, 2014 · d8c951c · d8c951c
1 parent 16a6ddc
commit d8c951c
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 5 deletions.
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
@@ -1283,12 +1283,28 @@ static int nfs_inode_attrs_need_update(const struct inode *inode, const struct n
 		((long)nfsi->attr_gencount - (long)nfs_read_attr_generation_counter() > 0);
 }
 
+/*
+ * Don't trust the change_attribute, mtime, ctime or size if
+ * a pnfs LAYOUTCOMMIT is outstanding
+ */
+static void nfs_inode_attrs_handle_layoutcommit(struct inode *inode,
+		struct nfs_fattr *fattr)
+{
+	if (pnfs_layoutcommit_outstanding(inode))
+		fattr->valid &= ~(NFS_ATTR_FATTR_CHANGE |
+				NFS_ATTR_FATTR_MTIME |
+				NFS_ATTR_FATTR_CTIME |
+				NFS_ATTR_FATTR_SIZE);
+}
+
 static int nfs_refresh_inode_locked(struct inode *inode, struct nfs_fattr *fattr)
 {
 	int ret;
 
 	trace_nfs_refresh_inode_enter(inode);
 
+	nfs_inode_attrs_handle_layoutcommit(inode, fattr);
+
 	if (nfs_inode_attrs_need_update(inode, fattr))
 		ret = nfs_update_inode(inode, fattr);
 	else
@@ -1518,8 +1534,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		if (new_isize != cur_isize) {
 			/* Do we perhaps have any outstanding writes, or has
 			 * the file grown beyond our last write? */
-			if ((nfsi->npages == 0 && !test_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags)) ||
-			     new_isize > cur_isize) {
+			if ((nfsi->npages == 0) || new_isize > cur_isize) {
 				i_size_write(inode, new_isize);
 				invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
 			}

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
@@ -7780,10 +7780,7 @@ nfs4_layoutcommit_done(struct rpc_task *task, void *calldata)
 	case -NFS4ERR_BADLAYOUT:     /* no layout */
 	case -NFS4ERR_GRACE:	    /* loca_recalim always false */
 		task->tk_status = 0;
-		break;
 	case 0:
-		nfs_post_op_update_inode_force_wcc(data->args.inode,
-						   data->res.fattr);
 		break;
 	default:
 		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
@@ -7798,6 +7795,8 @@ static void nfs4_layoutcommit_release(void *calldata)
 	struct nfs4_layoutcommit_data *data = calldata;
 
 	pnfs_cleanup_layoutcommit(data);
+	nfs_post_op_update_inode_force_wcc(data->args.inode,
+					   data->res.fattr);
 	put_rpccred(data->cred);
 	kfree(data);
 }

diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
@@ -359,6 +359,15 @@ pnfs_ld_layoutret_on_setattr(struct inode *inode)
 		PNFS_LAYOUTRET_ON_SETATTR;
 }
 
+static inline bool
+pnfs_layoutcommit_outstanding(struct inode *inode)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+
+	return test_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags) != 0 ||
+		test_bit(NFS_INO_LAYOUTCOMMITTING, &nfsi->flags) != 0;
+}
+
 static inline int pnfs_return_layout(struct inode *ino)
 {
 	struct nfs_inode *nfsi = NFS_I(ino);
@@ -515,6 +524,13 @@ pnfs_use_threshold(struct nfs4_threshold **dst, struct nfs4_threshold *src,
 	return false;
 }
 
+static inline bool
+pnfs_layoutcommit_outstanding(struct inode *inode)
+{
+	return false;
+}
+
+
 static inline struct nfs4_threshold *pnfs_mdsthreshold_alloc(void)
 {
 	return NULL;